Jun 2023 – Present · 2 yrs 9 mos · Bengaluru, Karnataka, India · On-site

• 3 products:
• 1. Confidential computing based anamoly/fraud detection in BFSI domain.
• 2. Bharat Vidya: Original Indic Voice data-sets for AI LLM pre-training
• 3. Yellow Nannies: Solving the consumer services problem in India. Creating an AI platform to connect urban masses to grassroots baby care nannies seeking employment and upskilling.
• Incubated at IIIT Bangalore, Nasscom, NVidia, STPI and Impact Orbit incubation program at NSRCEL, IIM Bangalore.
• Solving the most impactful consumer facing tasks in the entire globe, a 375 Billion USD market!
• tech@yellowsense.in || +91 - 9869.397.868

Feb 2021 – Jun 2023 · 2 yrs 4 mos · Bengaluru, Karnataka, India · On-site

• Managed following projects as part of M365 admin portal experience. Intent is to improvise the admin experience for tenant admins of enterprise customers like Accenture; they are using M365 products like Teams, OneDrive, SharePoint etc. (all projects part of E+D organization at IDC):
• 1. Teams Admin Center / Teams Powershell Module (TPM)
• Tenant admins of large employee strength enterprise customers need more optimized solutions for managing Teams usage by their employees. This was part of KR 'SfBO Decoupling' in IC3 India org.
• 2. OneDrive SharePoint RaaS (Repository as a Service) Admin
• For third party ISVs like Avepoint and Nintex - so that their enterprise customers like Walmart can better manage their usage of M365 storage.
• 3. Quota management in Share Point Online (SPO):
• Part of COGS savings for tenant admins.

Sep 2020 – Feb 2021 · 5 mos · Bengaluru, Karnataka, India · Remote

• Sep-Nov 2020: e-learning startup, launched by Trinamool Congress MP Shri Derek O'Brien.
• Developed a ed-tech platform which hones the real-life skills of students.
• This is very different from the regular curriculum applications which focus only on academic. Our platform focussed on personality development.
• Dec '20 - Feb '21: Geniemode.
• Built tech for B2B apparel sourcing domain industry.

Jan 2020 – Aug 2020 · 7 mos · Greater Bengaluru Area

• Reporting to the CTO, I head the entire engineering of Bangalore team of MagicBricks (Times of India group). This is a new initiative from Magicbricks - we have opened a new center in Bangalore called BIC.
• We are a part of Times group - Bennett, Coleman & Company Limited (BCCL).
• Key person in the architecture and entire development process (SDLC) of the module named MagicBricks "Commercials".

Jul 2019 – Jan 2020 · 6 mos · Greater Bengaluru Area

• Reporting to VP SuccessFactors, led technical architecture of Instance Refresh Tool (IRT) as part of SFSF.
• We used Docker Containers and Gradle Build automation (deployment) system on the top of HANA in-memory databases. Integrated SonarQube for automated static code analysis for our repository - Tenant Refresh module and propagated the usage of the corresponding Eclipse plugin- SonarLint for optimization in the software development best practices.
• https://blogs.sap.com/2019/07/16/new-sf-bizx-instance-refresh-process-tool/
• My group is a part of cloud based HCM (Human Capital Management) which got transformed into Human Experience Management (HXM).
• https://www.sap.com/india/products/human-resources-hcm/human-experience.html

Aug 2018 – Mar 2019 · 7 mos · Bengaluru Area, India

• Making travel cards online using cloud infrastructure:
• Created scalable, heavy-load systems, solving travel payments issues using fault-tolerant systems. Led the project architecture with a team of 14 people across Europe and India.
• Visited company headquarters at Eindhoven, Netherlands and delivered talk on mWallet project.
• Successfully released multiple versions to customers such as Mont Blanc and Sony.

Aug 2016 – Apr 2018 · 1 yr 8 mos · Greater Hyderabad Area

• Built the portal Finance Data Insights Portal (FinDI), to provide a single point of entry for reporting requirements. From this portal, end-users can search reports across reporting platforms like QuickSight, FACTBI, DARTBI and Grasshopper.
• Also part of big data analytics platform (Consumer Knowledge Platform-CKP) to support the analytics that required complex processing over large (200+ TB) amount of data with operations like: ingestion, transformation, feature-generation, model building, model scoring and publishing. The data was stored on AWS stores like S3 and processed using Big Data infrastructure like EMR & Hadoop Clusters. CKP provided a set of integrated abstractions using its metadata building blocks to create an extensible, loosely coupled data processing platform.
• Published the R&D paper at Amazon Machine Learning Conference (AMLC 2017) as a lead author at Seattle, USA. Tech – Java, Python, Redshift & DynamoDB Databases.

Jul 2015 – Dec 2015 · 5 mos · Pune/Pimpri-Chinchwad Area

• Key responsible for the entire architecture of the flagship product: DM-Active.
• Handles hadoop, map-reduce related issues.
• Disaster Recovery (DR) in Distributed Systems is one of the key contributions at DataMetica.

Nov 2014 – May 2015 · 6 mos · Pune/Pimpri-Chinchwad Area

• Innovation in the flagship R&D team: CREST at KPIT.

Sep 2012 – Jan 2013 · 4 mos · Gurugram, Haryana, India

• Optimizations in query response framework for organization's flagship platform implementing real-time data analytics.

Jul 2011 – Jul 2012 · 1 yr · Pune/Pimpri-Chinchwad Area

• Implemented new features/modues such as "customized offers" admin console, fixing issues in various modules such as knowledge base, system console.
• Successfully headed the responsibility of releasing the flagship product hot-release V-10.0.3.

Jun 2010 – Jul 2011 · 1 yr 1 mo · Mumbai Metropolitan Region

• 1. URL Classifier based on Cosine Similarity:
• URL is categorized based on the comparison between the term vectors of the url and that of the categories. A term vector for a keyword is a hashtable which contains the words related to that keyword and corresponding scores.
• 2. Context Classifier - Cosine Similarity
• For a given url, the content is first parsed and the main concepts are computed. For each concept, the related term vector is drawn from MS SQL database. Then the term vector of each of these related keywords and that of the url are compared. Cosine value is computed and the top 10 keywords are shown for each url context.

May 2007 – Jul 2007 · 2 mos · Greater Paris Metropolitan Region

• Policy Iteration Algorithms. Summer internship at INRIA Rocquencourt, Paris.